Borderline Personality Disorder (BPD) is a serious mental illness, associated with severe functional impairment and significant costs to society. Two significant problems impede successful diagnosis and treatment of these debilitating disorders. First is an almost complete lack of direct, empirical knowledge about actual symptom expression, frequency, or severity. Clinical diagnosis relies on comparisons of individual's symptom levels to normal frequencies and severities, but there is no direct, empirical knowledge about actual levels of symptom expression in either a normal or abnormal population. This limitation in data availability to the clinician makes it difficult to determine whether a given individual's symptoms qualify for diagnosis, are being effectively treated, or are reduced to healthy levels. A second significant problem is that there is little information about the processes underlying BPD and associated problems. Without knowledge of the underlying causal mechanisms, it is difficult to optimize effective treatment of BPD. The following research proposal is tailored to simultaneously address both of these significant problems. In R01 MH70571, Stage 1, an innovative approach that is particularly suited to solving these problems was developed and successfully applied to the study of normal personality. This approach is characterized by two hallmarks: (1) measuring actual behavior to describe and explain personality, rather than relying on retrospective questionnaires; and (2) leveraging variability in behavior to identify the mechanisms underlying personality. Stage 1 verified the usefulness of this approach for arriving at new insights about personality and for explaining the mechanisms underlying personality. Stage 2 will use this approach to accomplish the same goals for BPD. Stage 2 will measure actual symptoms rather than rely on retrospective reports, and will leverage variability in those symptoms to test the mechanisms underlying BPD. Aim 1 is to obtain direct, empirical accounts of symptom frequencies, severities, and co-occurrences, to aid diagnosis, treatment, and termination decisions. Aim 2 is to propose and test several theoretical mechanisms for BPD. Aim 3 is to investigate the role of interpersonal perception processes in stressors and symptoms. Aim 4 is to chart trajectories and transactions of symptom frequencies, severities, and contingencies. Borderline Personality Disorder, like normal personality, is often considered chronic and untreatable. However, this is a serious mental illness, associated with severe functional impairment, high health-care costs, and significant suffering. The key to finding treatments and reducing these costs is to find an approach that unlocks the mechanism underlying the disorder. Stage 1 demonstrated such an approach for normal personality; Stage 2 will do the same for borderline personality disorder.